A Fault Diagnosis Method for INS/DVL/USBL Integrated Navigation System Based on Support Vector Regression

Abstract

For the problem of fault subsystem identification in INS/DVL/USBL integrated navigation system, a fault diagnosis method based on SVR for INS/DVL/USBL integrated navigation system is proposed to improve the reliability of underwater robot integrated navigation system. The method firstly realizes the fault diagnosis of the integrated navigation system by the residual test method, but the traditional residual test method can only detect the fault and cannot accurately identify the fault subsystem. Therefore, a regression prediction model based on support vector machine is constructed to predict the state of the inertial navigation system. The fault diagnosis of the inertial navigation is assisted according to the difference between the output of the system model and the output of the prediction model, so as to identify the fault source of the system. Simulation experiments show that the method can diagnose the fault subsystem quickly and accurately. Through fault isolation and system reconstruction, the accuracy of the navigation system can be guaranteed, and the reliability and anti-interference of the integrated navigation system can be improved.